){kind=logo}
As AI technology continues to mature, in-house legal teams face novel challenges at the intersection of artificial intelligence and intellectual property. At Loeb & Loeb's 2026 Chicago AI Summit on July 14, in-house counsel from a wide range of industries convened for the "AI Issues with IP" roundtable to discuss the AI-related IP issues they encounter daily.
Here are some key takeaways from the discussion:
- Ownership of AI-generated content remains a central question. Participants highlighted the tension between leveraging AI to produce content quickly and cost-effectively while facing genuine uncertainty about whether the output is protectable. Where AI involvement is significant, copyright and other IP claims may be weakened or entirely unavailable, leaving companies with assets they cannot fully protect or enforce.
- Copyright infringement risks are a growing concern. Equally pressing was the risk that AI outputs incorporate or closely resemble third-party protected works. Rights clearance becomes even more complex when content is repurposed through AI tools, particularly for commercial use.
- Likeness and publicity rights risks are top of mind. Several participants raised concerns about AI being used to recreate or modify a person's image, voice or identity, implicating publicity rights and related claims. The group discussed how companies are approaching consent frameworks and whether pre-AI contractual terms adequately cover AI-generated uses.
- Contracts and liability allocation are struggling to keep up. Traditional indemnification and warranty frameworks are often ill-suited to address how AI introduces legal exposure. Small changes in prompts can lead to significantly different legal outcomes, making it difficult to define contractual obligations with precision.
- Governance gaps and documentation deficiencies persist. Insufficient documentation of prompts and AI workflows makes ownership and compliance harder to prove after the fact. Employee non-compliance with AI policies (whether through ignorance or convenience) remains a persistent problem, compounded by reliance on unvetted AI models trained on unknown or potentially problematic datasets.
- Downstream use and the "embedded AI" question. The group discussed the growing risk of unauthorized downstream use of company IP by third parties leveraging AI tools to remix or repurpose proprietary content. Participants also explored whether embedded AI features in enterprise software create different risk profiles than standalone AI applications, and whether existing governance frameworks adequately address both.
